The high variability in patient skeletal response to drug treatments for osteoporosis presupposes an important clinical question: which patients are likely to benefit from a particular intervention? A few studies have examined the capacity of baseline characteristics to predict skeletal response to therapy, but these studies have been limited to alendronate and have focused on areal bone mineral density (aBMD) changes as an endpoint. Although aBMD is associated with incident hip fracture, it is simply a summary parameter of integrated bone mass, and provides no information about changes in whole bone structure and strength, factors that would lend direct mechanistic insight into differences between therapies, and between individuals. The overarching goal of our proposed study is to understand the factors underlying the variability between individuals and treatments of the effect of anti-resorptive (alendronate and zolendronate) and anabolic (parathyroid hormone-PTH) treatments on proximal femoral whole bone strength. We will accomplish this reanalyzing anonymized QCT scans obtained in three different clinical drug studies: HORIZON (zolendronate), PaTH (PTH, alendronate and combination therapy) and OPTAMISE (PTH after cessation of anti-resropritve treatment). As an endpoint measure, we will employ whole-bone strength assessed by volumetric quantitative computed tomography (vQCT)-based subject specific finite element modeling (FEM). In order to visualize, on a voxel-by voxel basis, the BMD and geometric changes that underlie changes in strength, we will employ voxel-based morphometry (VBM), a neuro-image processing technique adapted by our laboratory to the study of bone. We will use VBM to map changes in bone structure, to relate strength changes to baseline structural features, and we will integrate VBM with the stress, strain fields generated by the FEM in order to relate changes in strength and structure to the distribution of stress, strain and strain energy density at baseline. We believe that our experimental approach will provide novel scientific information along with a set of results that can be translated to clinical practic.